The computations performed by the neocortex result from the activity of neural circuits composed of glutamatergic excitatory and GABAergic inhibitory neurons. Although representing only 15–30% of all cortical neurons, inhibitory neurons profoundly influence cortical computations and cortical dynamics. For example, they influence how excitatory neurons integrate information, shape neuronal tuning properties, modulate neuronal responses based on sensory context and behavioral state, and maintain an appropriate dynamic range of cortical excitation (Ferster and Miller, 2000; Shapley et al., 2003; Tremblay et al., 2016). These various functions of inhibition are thought to be mediated by different inhibitory neuron types, of which a large diversity has been identified in several species, each having distinct chemical, electrophysiological, and morphological properties (Ascoli et al., 2008; Burkhalter, 2008; Kubota et al., 2016).

In mouse cortex, the expression of specific molecular markers identifies three major, largely non-overlapping classes of inhibitory neurons: parvalbumin- (PV), somatostatin- (SOM), and serotonin receptor (5HT3aR, a larger class which includes vasoactive intestinal peptide or VIP cells)-expressing neurons (Xu et al., 2010; Rudy et al., 2011). The creation of mouse lines selectively expressing Cre-recombinase in specific inhibitory neuron classes has led to a multitude of studies on the connectivity and function of each class (Tremblay et al., 2016; Wood et al., 2017; Shin and Adesnik, 2023). Distinct patterns of connectivity and function specific to each inhibitory neuron class are emerging from these mouse studies, but it remains unknown whether insights gained from mouse apply to inhibitory neurons in higher species such as primates. Understanding cortical inhibitory neuron function in the primate is critical for understanding cortical function and dysfunction in the model system closest to humans, where cortical inhibitory neuron dysfunction has been implicated in many neurological and psychiatric disorders, such as epilepsy, schizophrenia, and Alzheimer’s disease (Cheah et al., 2012; Verret et al., 2012; Mukherjee et al., 2019).

A major impediment to studying inhibitory neuron function in primates has been the lack of tools for cell-type-specific expression of transgenes in this species. However, recent advances in the application to primates of cell-type-specific viral technology are beginning to enable studies of inhibitory neuron types in primate cortex. In particular, two recent studies have developed specific promoters or enhancers that restrict transgene expression from recombinant adeno-associated viral vectors (AAV) to GABAergic neurons, specifically the mDlx enhancer (Dimidschstein et al., 2016) and the h56D promoter (Mehta et al., 2019), in both rodents and primates. Viral strategies to restrict gene expression to PV neurons have also been recently developed (Mehta et al., 2019; Vormstein-Schneider et al., 2020; Mich et al., 2021).

To facilitate the application of these inhibitory-neuron specific viral vectors to studies of the primate cortex, we have performed a thorough validation and characterization of the laminar expression of reporter proteins mediated by several enhancer/promoter-specific AAVs. Here, we report results from the two vector types that have shown the greatest specificity of transgene expression in marmoset primary visual cortex (V1); specifically, we have tested three serotypes of the h56D promoter-AAV that restricts gene expression to GABAergic neurons (Mehta et al., 2019), and one serotype of the S5E2 enhancer-AAV that restricts gene expression to PV cells (Vormstein-Schneider et al., 2020). Using injections of these viral vectors in marmoset V1, combined with immunohistochemical identification of GABA and PV neurons, we find that the laminar distribution of reporter protein expression mediated by the GABA- and PV-enhancer AAVs validated in this study resembles the laminar distribution of GABA-immunoreactive (GABA+) and PV-immunoreactive (PV+) cells, respectively, in marmoset V1. Reporter protein expression mediated by the h56D-AAV is specific and robust, but the degree of specificity and coverage depended on serotype and cortical layer. We found that about 92% of PV cells in marmoset V1 are GABA+, and reporter protein expression mediated by the S5E2-AAV shows up to 98% specificity and 86–90% coverage, depending on layer. We conclude that these viral vectors offer the possibility of studying GABAergic and PV neuron connectivity and function in primate cortex.

We validated three serotypes (1,7,9) of a pAAV-h56D-tdTomato (Mehta et al., 2019) and the AAV-PHP.eB-S5E2.tdTomato (Vormstein-Schneider et al., 2020). We report results from 10 viral injections, of which 3 injections of AAV-h56D-tdTomato, and 7 injections of AAV-PHP.eB-S5E2.tdTomato, made in 4 marmoset monkeys (see ‘Materials and methods’ and Supplementary file 1). Tissue sections through V1 were double immunoreacted for GABA and PV and imaged on a fluorescent microscope. We quantified the laminar distribution of viral-induced tdTomato (tdT) expression as well as of GABA+ and PV+ cells revealed by immunohistochemistry (IHC), and counted double- and triple-labeled cells to determine the specificity and coverage of viral-induced tdT expression across marmoset V1 layers (see ‘Materials and methods’).

V1 laminar distribution of GABA+ and PV+ neurons

We first determined the V1 laminar distribution of GABA+ and PV+ neurons identified by IHC (Figure 1). To this goal, in each section used for analysis, we counted GABA+ and PV+ cells within 2 × 100-µm-wide regions of interest (ROIs) spanning all layers in dorsal V1 anterior to the posterior pole, for a total of six ROIs in three tissue sections selected randomly. Cortical layer boundaries were determined using DAPI and/or PV staining (Figure 1B and C), as we found that PV-IHC reveals laminar boundaries consistent with those defined by DAPI.

Figure 1

Download asset Open asset

The laminar distribution of GABA+ and PV+ cells was quantified as percent of total GABA+ or PV+ cells (Figure 2A), as well as cell density (number of cells per unit area; Figure 2B), in each cortical layer. Both GABA+ and PV+ cell percent and density peaked in layers (L) 2/3 and 4C. There was no significant difference in the percent or density of GABA+ cells in L2/3 (33% ± 2, and 1294 cells/mm² ± 74.4, respectively) vs. L4C (33.4% ± 3.1, and 1052 cells/mm² ± 42.2, respectively), as determined by a Bonferroni-corrected Kruskal–Wallis test (p=1.00, n = 6 ROIs across three sections, L2/3 vs. L4C in Figure 2A) or by a Bonferroni-corrected ANOVA (p=1.00, n = 6 ROIs across three sections, L2/3 vs. L4C in Figure 2B). Similarly, the percent and density of PV+ cells in L4C (42.3% ± 3.6, and 888 cells/mm² ± 38.8, respectively) were not significantly different from those in L2/3 (31.3% ± 2.5 and 812 cells/mm² ± 65, respectively; p=1.00 for both comparisons, n = 6 ROIs). However, the density of PV+ cells in L4C was significantly higher than that in all remaining layers (p=0.028005 for 4C vs. 4A/B, and <0.001 for 4C vs. all other layers; ANOVA with Bonferroni correction; n = 6 ROIs), and L2/3 PV+ density was significantly higher than L1, 5 and 6 (p=5.93E^–11 for L2/3 vs. L1 and p=0.00037 L2/3 vs. L6, p=0.002 for L2/3 vs. L5, n = 6 ROIs), but not L4A/B (although the percent of PV+ cells in L2/3 was significantly higher than that in L4A/B; p=0.029 Bonferroni-corrected Kruskal–Wallis test, n = 6 ROIs). GABA+ cell density in L2/3 was significantly higher than that in all other layers except L4C (p=0.028 vs. L4A/B and p=0.013 vs. L6, p=0.007 vs. L1 and p=0.005 vs. L5). GABA density in L4C did not differ from any other layers, but the percent of GABA+ cells in L4C was significantly higher than that in L1 (p=0.009) and 4A/B (p=0.000022). As expected, within each layer, GABA+ cell density was significantly higher than PV+ cell density (p<0.05, one-sided t-test for equality of means, n = 6 ROIs).

Figure 2

Download asset Open asset

We compared the laminar distributions of GABA+ and PV+ cell density in marmoset V1 with previously published distributions of these two cell markers in mouse V1 (Xu et al., 2010). In marmoset V1, there is an overall higher density of PV+ cells than in mouse V1, with density peaking in L2/3 and 4C. In contrast, PV+ cell density in mouse V1 peaks in L4 and 5, and density in L2/3 is lower than that in all other layers (Figure 2C). GABA+ cell density in marmoset V1 peaks in L2/3 followed by 4C, whereas it peaks in L4 and 5 with a smaller third peak in L2/3 in mouse V1 (Figure 2D).

Counts of cells double labeled for GABA+ and PV+ revealed that 92.3% ± 1.9 of PV+ cells across all layers were GABA+, and that PV+ cells represent on average 61.4% ± 2.7 of all GABA+ cells, ranging from 54.5% ± 3.7 in L6 to 78.5% ± 5.6 in L4C (Figure 2E). This differs from mouse V1 in which PV+ cells represent about 40% of all GABA cells (Xu et al., 2010).

Laminar specificity and coverage of GABA-specific AAV-h56D

Figure 3 shows tdT expression at three injection sites, each of a different serotype (9,7,1) of the GABA-specific AAV-h56D-tdT. Identical injection volumes of each serotype, delivered at three different cortical depths (see ‘Materials and methods’), resulted in viral expression regions that differed in both size as well as laminar distribution, suggesting the different serotypes may have different capacity of infecting cortical neurons and layers. The AAV7 injection resulted in the smallest expression region, which additionally was biased to the superficial and deep layers, with only a few cells expressing tdT in the middle layers (Figure 3B). AAV9 (Figure 3A) and AAV1 (Figure 3C) resulted in larger expression regions, which involved all cortical layers. Given identical volumes and titers used for the AAV9 and AAV7 injections (injected volume of the AAV1 was the same but titer was higher; see Supplementary file 1), as well as identical post-injection survival times for all three serotypes, the differences in the size of the expression region are likely due to different tropism and/or viral spread of the different serotypes. However, given that we only made a single injection per serotype, we cannot exclude that other factors may have contributed to the reduced spread of the AAV7.

Figure 3

Download asset Open asset

In Figure 4, we report quantitative counts for each serotype. Below we describe these results, but acknowledge that the differences we observed between serotypes need to be interpreted with caution, given they are based on a single injection per serotype. Figure 4A compares quantitatively tdT expression obtained with each viral serotype, quantified as the percent of total tdT+ cells in each layer for each serotype, with the percent laminar distribution of GABA+ cells identified by IHC. Serotypes 9 and 1 overall showed similar laminar distribution as GABA+ IHC, the percent of tdT+ cells peaking in L2/3 and 4C, suggesting good specificity of viral infection (the laminar distributions of AAV9- and AAV1-induced tdT+ cells were not significantly different from the GABA+ cell distribution; p>0.05 for all comparisons, Bonferroni-corrected independent-samples median test, n = 4–6 ROIs). In contrast, due to the lower capacity of AAV7 to infect the mid-layers, AAV7-induced tdT expression was relatively higher in L2/3 compared to GABA+ expression, approaching statistical significance (p=0.059; Bonferroni-corrected independent-samples median test, n = 4 AAV7 ROIs and 6 GABA-IHC ROIs). The percent of AAV7-infected cells was also significantly higher than the percent of AAV9- and/or AAV1-infected cells in L2/3 (p=0.028 for both comparisons) and in L6 (p=0.028 for AAV7 vs. AAV1), and significantly lower than the percent of AAV9- and AAV1-infected cells in L4C (p=0.028 for both comparisons; Bonferroni-corrected independent-samples median test, n = 4 ROIs).

Figure 4

Download asset Open asset

To quantify the specificity of tdT expression induced by each serotype, that is, the accuracy in inducing tdT expression selectively in GABA cells, for each serotype separately we measured the percent of tdT-expressing cells that colocalized with GABA expression revealed by IHC (Figure 4B). Overall, across all layers, AAV9 showed the highest specificity (82.3% ± 1.1) followed by AAV7 (79.2% ± 5.4), and AAV1 (75.3% ± 2.6), and there was a statistically significant difference in overall specificity between AAV9 and AAV1 (p=0.014; Bonferroni-corrected independent-samples median test, n = 4 ROIs for each serotype). The specificity of AAV9 did not differ significantly across layers, ranging from 80.4% ± 2.1 in L4C to 93.8% ± 6.3 in L4A/B. In contrast, specificity for the other two serotypes varied by layer. AAV7 showed highest specificity in L1 (100% but there were only two tdT+ cells in this layer), L4C (90.1% ± 5.9) and L6 (87.1% ± 8.1), and lowest in L4A/B (50% ± 35.4). AAV1 specificity was highest in L6 (85.4% ± 8.8) and L4C (81.6% ± 1.4) and lowest in L4A/B (38.3% ± 21.7). There was a tendency for AAV9 to be more specific than AAV1 in L4A/B (93.8% ± 6.3 vs. 38.3% ± 21.7) and L5 (88.8% ± 6.6 vs. 59.4% ± 8.3) but these differences did not reach statistical significance.

To quantify the efficiency of the virus in inducing tdT expression in GABA cells, for each serotype separately we measured the viral coverage as the percent of GABA+ cells within the viral injection site that colocalized with tdT expression (Figure 4C). Overall, across all layers, AAV9 and AAV1 showed significantly higher coverage (66.1% ± 3.9 and 64.9% ± 3.7) than AAV7, which showed much lower coverage values (34% ± 5.6; p=0.014 for both comparisons; Bonferroni-corrected independent-samples median test, n = 4 ROIs across two sections for each serotype). AAV9 and AAV1 coverage was similar across layers, and both showed slightly higher coverage in superficial (AAV9: 67.4% ± 2.5; AAV1: 69% ± 6.5) and middle layers (AAV9: 78.5% ± 9.1; AAV1: 76.9% ± 7.4), compared to deep layers (AAV9: 50–55%, AAV1: 44–60%). Instead, AAV7 showed very low coverage values in L4A/B (8.3% ± 8.3) and L4C (14.4% ± 6.7) and highest values in L6 (47.9% ± 4.3) followed by L2/3 (44.6% ± 9). AAV7 coverage was significantly lower than AAV9 coverage in L2/3 (p=0.014), L4A/B (p=0.014), and L4C (p=0.014), and was significantly lower than AAV1 in L4C (p=0.014; Bonferroni-corrected independent-samples median test, n = 4 ROIs for each serotype). Thus, our results suggest that AAV9 is the serotype of choice for marmoset studies of GABAergic neurons requiring highest specificity and coverage across all layers, but AAV7 may be a better choice for studies intending to restrict transgene expression to L6 or L2/3 GABA cells with good specificity.

Laminar specificity and coverage of PV-specific AAV-PHP.eB-S5E2

We assessed the laminar specificity and coverage of the AAV-PHP.eB-S5E2-tdT following injections of different viral volumes ranging from 90 nl to 585 nl (see Supplementary file 1). Figure 5 shows fluorescent images of tdT expression at the site of viral injection for an example 105 nl injection (Figure 5A) and an example 315 nl injection (Figure 5B).

Figure 5

Download asset Open asset

Figure 6A compares tdT expression resulting from injections of different volumes, quantified as the percent of total tdT+ cells in each layer for each volume, with the percent laminar distribution of PV+ cells identified by IHC. Cell counts from injections of 315–585 nl volumes were pooled as injections ≥ 315 nl produced similar results (see Supplementary file 2). We found that the distribution of tdT expression resulting from all injection volumes did not differ significantly from the distribution of PV+ IHC, all distributions similarly peaking in L2/3 and 4C (p>0.1 for all comparisons; Bonferroni-corrected Kruskal–Wallis test; n = 8–12 PV-AAV ROIs, and 6 PV-IHC ROIs), suggesting good viral specificity.

Figure 6

Download asset Open asset

The specificity of tdT expression induced by different injection volumes is quantified in Figure 6B, separately for three groups of injection volumes: group 1 = 90–105 nl, group 2 = 180 nl, group 3 = 315–585 nl. Results from individual injection cases are reported in Supplementary file 2. The degree of viral specificity was high at all volumes, but depended slightly on injection volume. Overall, across all layers, group 2 (180 nl) showed the highest specificity (94.7% ± 1.6), which differed significantly from the specificity of group 3 volumes (≥315 nl; 82% ± 3.2; p=0.01, Bonferroni-corrected Kruskal–Wallis test, n = 8–12 ROIs). Specificity was similar across layers for all volumes, but volumes ≥ 315 nl resulted in slightly lower specificity than smaller volumes in L4C (76.6% ± 5.6 for >315 nl volumes vs. 95.2% ± 1.7 and 94.9% ± 3 for 90–105 nl and 180 nl volumes, respectively) and L5 (80.1% ± 5.8 for ≥315 nl vs. 97.9% ± 2.1 and 97% ± 1.9 for 90–105 and 180 nl, respectively), and these differences in L5 were statistically significant (p=0.013 and 0.005; Bonferroni-corrected Kruskal–Wallis test; n = 8–12 ROIs).

The viral coverage resulting from each injection volume is quantified in Figure 6C separately for the three different volume groups and shown for each individual injection case in Supplementary file 2. Coverage of the AAV-PHP.eB-S5E2-tdT was high, did not depend on injection volume, and it was similar across layers for all volumes. Overall, across all layers coverage ranged from 78% ± 1.9 to 81.6% ± 1.8.

Reduced GABA and PV immunoreactivity at the viral injection site

Qualitative observations of tissue sections seemed to indicate slightly reduced expression of both GABA and PV immunoreactivity at the viral injection sites, extending beyond the borders of the injection core (Figure 7—figure supplement 1). To quantify this observation, we counted GABA+ and PV+ cells at the site of the viral injections (n = 12 ROIs across 6 sections for AAV-h56D injection sites [pooled across serotypes], and 28 ROIs across 14 sections for AAV-S5E2 injection sites) and at sites located several millimeters beyond the viral injection borders (n = 6 ROIs across three sections). We found that both the number and density of GABA+ and PV+ cells were reduced across all layers at the site of the AAV-h56D (Figure 7A–D) and AAV-S5E2 (Figure 7E–H) injections compared to control tissue away from the injection sites. The magnitude of the reduction in immunoreactivity depended on the viral type, with the AAV-h56D virus inducing an overall greater and more significant reduction in GABA immunoreactivity (28.1% and 21.5% reduction in mean GABA+ cell number and density across all layers, respectively, p=0.024 in Figure 7A and p=0.013 in Figure 7B, Mann–Whitney U test) than in PV immunoreactivity (20.5 and 10.2% reduction in mean PV cell number and density across all layers, respectively; p=0.041 in Figure 7C and p=0.125 in Figure 7D, Mann–Whitney U test) and vice versa for the AAV-S5E2 virus, which reduced PV immunoreactivity (33.3% and 25.4% reduction in mean PV+ cell number and density across all layers, respectively, p<0.001 in Figure 7G and p=0.013 in Figure 7H) more than GABA immunoreactivity (27.4% and 20.2% reduction in mean GABA+ cell number and density across all layers, respectively, p=0.005 in Figure 7E and p=0.042 in Figure 7F). The reduced GABA and PV immunoreactivity caused by the viruses imply that the specificity of the viruses we have validated in this study is likely higher than estimated in Figures 4 and 6. Moreover, reduced GABA and PV immunoreactivity could at least partly underlie the apparent reduction in specificity observed for larger PV-AAV injection volumes (see ‘Discussion’).

Figure 7 with 1 supplement see all

Download asset Open asset

Understanding the connectivity and function of inhibitory neurons and their subtypes in primate cortex requires the development of viral tools that allow for specific and robust transgene expression in these cell types. Recently, several enhancer and promoter elements have been identified that allow to selectively and efficiently restrict gene expression from AAVs to GABAergic neurons and their subtypes across several species, but a thorough validation and characterization of these enhancer-AAVs in primate cortex is lacking. In particular, previous studies have not characterized the specificity and coverage of these vectors across cortical layers. In this study, we have characterized two main enhancer-AAV vectors designed to restrict expression to GABAergic cells or their PV subtypes, which show high specificity and coverage in marmoset V1. Specifically, we have shown that the GABA-specific AAV9-h56D (Mehta et al., 2019) induces transgene expression in GABAergic cells with up to 91–94% specificity and 80% coverage, depending on layer, and the PV-specific AAV-PHP.eB-S5E2 (Vormstein-Schneider et al., 2020) induces transgene expression in PV cells with up to 98% specificity and 86–90% coverage, also depending on layer. We conclude that these two viral vector types provide useful tools to study inhibitory neuron connectivity and function in primate cortex.

Many recent studies have investigated the connectivity and function of distinct classes of inhibitory neurons in mouse V1 and other cortical areas (Tremblay et al., 2016; Wood et al., 2017; Shin and Adesnik, 2023). In contrast, similar studies in the primate have been missing due to the lack of tools to selectively express transgenes in specific cell types and the difficulty of performing genetic manipulation in this species. It is important to study inhibitory neuron function in the primate because it is unclear whether findings in mice apply to higher species, and inhibitory neuron dysfunction in humans has been implicated in several neurological and psychiatric disorders (Marín, 2012; Goldberg and Coulter, 2013; Lewis, 2014). While the basic inhibitory neuron subtypes seem to exist across most mammalian species studied (DeFelipe, 2002), species differences may exist, particularly given the evolutionary distance between mouse and primate. Indeed, species differences have been reported in marker expression patterns (Hof et al., 1999), in the proportion of cortical GABAergic neurons (24–30% in primates vs. 15% in rodents) (Hendry et al., 1987; Beaulieu, 1993), in the proportion of PV neurons (74% in macaque V1 vs. 30–40% in mouse) (Van Brederode et al., 1990; Defelipe et al., 1999; Xu et al., 2010), and in the abundance of the various subtypes (Krienen et al., 2020). Here, we found that PV cells in marmoset V1 across all layers represent on average 61% of all GABAergic cells, and up to 79% in V1 L4C. These percentages are lower than previously reported for macaque V1 by Van Brederode et al., 1990 (74% across all layers and nearly 100% in L4C), but higher than recently reported by Kelly et al., 2019 (52% across all V1 layers, up to 80% in L4C). We also found differences in the V1 laminar expression of both GABA+ and PV+ cells between mouse and marmoset. Specifically, GABA+ and PV+ expression peaks in L2/3 and 4C in marmoset V1, but in L4 and 5 in mouse V1. Similar differences between mouse and primate V1 in the laminar distribution of PV cells were reported previously (Kooijmans et al., 2020; Medalla et al., 2023). Our results on the laminar distribution of PV and GABA immunoreactivity are consistent with previous qualitative and quantitative studies in macaque V1 (Hendry et al., 1989; Blümcke et al., 1990; Defelipe et al., 1999; Disney and Aoki, 2008; Kelly et al., 2019; Kooijmans et al., 2020; Medalla et al., 2023), and with a quantitative study in marmoset V1 (Goodchild and Martin, 1998).

We compared laminar distribution, specificity, and coverage of three different serotypes of the AAV-h56D vector. Serotypes 9 and 7 showed slightly greater specificity than serotype 1, and the specificity of AAV9 was more consistent across layers than the specificity of serotypes 7 and 1, which instead varied somewhat across layers. Serotypes 9 and 1 showed greater coverage than serotype 7. Thus, serotype 9 may be a better choice when high specificity and coverage across all layers are required. Serotype 7, instead, showed high specificity (80%) but low coverage (34%), except in layer 6 (48%) and L2/3 (45%); therefore, this serotype may be desirable to restrict transgene expression to L6 or 2/3 GABAergic cells. We note that these differences among serotypes should be interpreted with caution as they are based on a single injection per serotype. Despite this, our results demonstrate sufficiently high efficiency and specificity of transgene expression in GABA cells using the h56D promoter, at least with two of the three AAV serotypes we tested, warranting their use in the non-human primate.

We compared laminar distribution, specificity, and coverage resulting from different volume injections of the AAV-PHP.eB-S5E2 vector. Injections of 180 nl volume resulted in higher specificity (95% across layers) and coverage (81% across all layers) than obtained with injection volumes equal to or larger than 315 nl (specificity 82% and coverage 78% across all layers), although coverage did not differ significantly across volumes. This mild dose-dependent alteration of specificity could depend on some off-target expression reaching above detection levels at higher doses. Thus, injection volumes of 150–300 nl are recommended for studying PV neuron function and connectivity using this viral vector. Alternatively, or in addition, an apparent dose-dependent reduction in specificity may result from viral-induced suppression of PV immunoreactivity, which could be more pronounced for larger volume injections. Indeed, we found that both GABA- and PV-specific AAVs slightly, but significantly, reduced both GABA and PV immunoreactivity at the site of the viral injection, but GABA expression was more reduced at the AAV-h56D injection site, while reduction in PV expression was more marked at the AAV-S5E2 injection site. This reduction in GABA and PV immunoreactivity at the viral injected sites most likely affected our measurements of specificity, suggesting that specificity for the viruses tested in this study is even higher than revealed by our counts. However, this reduced immunoreactivity raises concerns about the virus or the high level of reporter protein possibly harming the cell physiology. Our data does not allow us to assess the origin of the reduced GABA and PV immunoreactivity. Qualitative observations did not reveal structural damage at the site of the viral injections to suggest cell death. Moreover, we have been able to record the electrophysiological responses of V1 neurons in which opsin protein expression was induced via injections of these viruses (Vafaei et al., 2024). Notably, viral-induced downregulation of gene expression in host cells, including of inhibitory neuron marker genes such as PV, has been documented for other viruses, such as rabies virus (Prosniak et al., 2001; Zhao et al., 2011; Patiño et al., 2022). As such, it is possible that subtle alteration of the cortical circuit upon parenchymal injection of viruses (including AAVs) leads to alteration of activity-dependent expression of PV and GABA.

Source data for the figures can be found on DRYAD at https://doi.org/10.5061/dryad.ht76hdrr3.

1. 1. Federer F
   2. Angelucci A

   (2024) Dryad Digital Repository

   Laminar specificity and coverage of viral-mediated gene expression restricted to GABAergic interneurons and their parvalbumin subclass in marmoset primary visual cortex.

   https://doi.org/10.5061/dryad.ht76hdrr3

1. 1. Ascoli GA
   2. Alonso-Nanclares L
   3. Anderson SA
   4. Barrionuevo G
   5. Benavides-Piccione R
   6. Burkhalter A
   7. Buzsáki G
   8. Cauli B
   9. Defelipe J
   10. Fairén A
   11. Feldmeyer D
   12. Fishell G
   13. Fregnac Y
   14. Freund TF
   15. Gardner D
   16. Gardner EP
   17. Goldberg JH
   18. Helmstaedter M
   19. Hestrin S
   20. Karube F
   21. Kisvárday ZF
   22. Lambolez B
   23. Lewis DA
   24. Marin O
   25. Markram H
   26. Muñoz A
   27. Packer A
   28. Petersen CCH
   29. Rockland KS
   30. Rossier J
   31. Rudy B
   32. Somogyi P
   33. Staiger JF
   34. Tamas G
   35. Thomson AM
   36. Toledo-Rodriguez M
   37. Wang Y
   38. West DC
   39. Yuste R
   40. Petilla Interneuron Nomenclature Group

   (2008) Petilla terminology: nomenclature of features of GABAergic interneurons of the cerebral cortex

   Nature Reviews. Neuroscience 9:557–568.

   https://doi.org/10.1038/nrn2402

   • PubMed
   • Google Scholar
2. 1. Aurnhammer C
   2. Haase M
   3. Muether N
   4. Hausl M
   5. Rauschhuber C
   6. Huber I
   7. Nitschko H
   8. Busch U
   9. Sing A
   10. Ehrhardt A
   11. Baiker A

   (2012) Universal real-time PCR for the detection and quantification of adeno-associated virus serotype 2-derived inverted terminal repeat sequences

   Human Gene Therapy Methods 23:18–28.

   https://doi.org/10.1089/hgtb.2011.034

   • PubMed
   • Google Scholar
3. 1. Beaulieu C

   (1993) Numerical data on neocortical neurons in adult rat, with special reference to the GABA population

   Brain Research 609:284–292.

   https://doi.org/10.1016/0006-8993(93)90884-p

   • PubMed
   • Google Scholar
4. 1. Blümcke I
   2. Hof PR
   3. Morrison JH
   4. Celio MR

   (1990) Distribution of parvalbumin immunoreactivity in the visual cortex of Old World monkeys and humans

   The Journal of Comparative Neurology 301:417–432.

   https://doi.org/10.1002/cne.903010307

   • PubMed
   • Google Scholar
5. 1. Burkhalter A

   (2008) Many specialists for suppressing cortical excitation

   Frontiers in Neuroscience 2:155–167.

   https://doi.org/10.3389/neuro.01.026.2008

   • PubMed
   • Google Scholar
6. 1. Cheah CS
   2. Yu FH
   3. Westenbroek RE
   4. Kalume FK
   5. Oakley JC
   6. Potter GB
   7. Rubenstein JL
   8. Catterall WA

   (2012) Specific deletion of NaV1.1 sodium channels in inhibitory interneurons causes seizures and premature death in a mouse model of Dravet syndrome

   PNAS 109:14646–14651.

   https://doi.org/10.1073/pnas.1211591109

   • PubMed
   • Google Scholar
7. 1. Defelipe J
   2. González-Albo MC
   3. Del Río MR
   4. Elston GN

   (1999) Distribution and patterns of connectivity of interneurons containing calbindin, calretinin, and parvalbumin in visual areas of the occipital and temporal lobes of the macaque monkey

   The Journal of Comparative Neurology 412:515–526.

   https://doi.org/10.1002/(sici)1096-9861(19990927)412:3<515::aid-cne10>3.0.co;2-1

   • PubMed
   • Google Scholar
8. 1. DeFelipe J

   (2002) Cortical interneurons: from Cajal to 2001

   Progress in Brain Research 136:215–238.

   https://doi.org/10.1016/s0079-6123(02)36019-9

   • PubMed
   • Google Scholar
9. 1. Dimidschstein J
   2. Chen Q
   3. Tremblay R
   4. Rogers SL
   5. Saldi G-A
   6. Guo L
   7. Xu Q
   8. Liu R
   9. Lu C
   10. Chu J
   11. Grimley JS
   12. Krostag A-R
   13. Kaykas A
   14. Avery MC
   15. Rashid MS
   16. Baek M
   17. Jacob AL
   18. Smith GB
   19. Wilson DE
   20. Kosche G
   21. Kruglikov I
   22. Rusielewicz T
   23. Kotak VC
   24. Mowery TM
   25. Anderson SA
   26. Callaway EM
   27. Dasen JS
   28. Fitzpatrick D
   29. Fossati V
   30. Long MA
   31. Noggle S
   32. Reynolds JH
   33. Sanes DH
   34. Rudy B
   35. Feng G
   36. Fishell G

   (2016) A viral strategy for targeting and manipulating interneurons across vertebrate species

   Nature Neuroscience 19:1743–1749.

   https://doi.org/10.1038/nn.4430

   • PubMed
   • Google Scholar
10. 1. Disney AA
    2. Aoki C

    (2008) Muscarinic acetylcholine receptors in macaque V1 are most frequently expressed by parvalbumin-immunoreactive neurons

    The Journal of Comparative Neurology 507:1748–1762.

    https://doi.org/10.1002/cne.21616

    • PubMed
    • Google Scholar
11. 1. Ferster D
    2. Miller KD

    (2000) Neural mechanisms of orientation selectivity in the visual cortex

    Annual Review of Neuroscience 23:441–471.

    https://doi.org/10.1146/annurev.neuro.23.1.441

    • PubMed
    • Google Scholar
12. 1. Goldberg EM
    2. Coulter DA

    (2013) Mechanisms of epileptogenesis: a convergence on neural circuit dysfunction

    Nature Reviews. Neuroscience 14:337–349.

    https://doi.org/10.1038/nrn3482

    • PubMed
    • Google Scholar
13. 1. Goodchild AK
    2. Martin PR

    (1998) The distribution of calcium-binding proteins in the lateral geniculate nucleus and visual cortex of a New World monkey, the marmoset, Callithrix jacchus

    Visual Neuroscience 15:625–642.

    https://doi.org/10.1017/s0952523898154044

    • PubMed
    • Google Scholar
14. 1. Grieger JC
    2. Snowdy S
    3. Samulski RJ

    (2006) Separate basic region motifs within the adeno-associated virus capsid proteins are essential for infectivity and assembly

    Journal of Virology 80:5199–5210.

    https://doi.org/10.1128/JVI.02723-05

    • PubMed
    • Google Scholar
15. 1. Hendry SH
    2. Schwark HD
    3. Jones EG
    4. Yan J

    (1987) Numbers and proportions of GABA-immunoreactive neurons in different areas of monkey cerebral cortex

    The Journal of Neuroscience 7:1503–1519.

    https://doi.org/10.1523/JNEUROSCI.07-05-01503.1987

    • PubMed
    • Google Scholar
16. 1. Hendry SH
    2. Jones EG
    3. Emson PC
    4. Lawson DE
    5. Heizmann CW
    6. Streit P

    (1989) Two classes of cortical GABA neurons defined by differential calcium binding protein immunoreactivities

    Experimental Brain Research 76:467–472.

    https://doi.org/10.1007/BF00247904

    • PubMed
    • Google Scholar
17. 1. Hof PR
    2. Glezer II
    3. Condé F
    4. Flagg RA
    5. Rubin MB
    6. Nimchinsky EA
    7. Vogt Weisenhorn DM

    (1999) Cellular distribution of the calcium-binding proteins parvalbumin, calbindin, and calretinin in the neocortex of mammals: phylogenetic and developmental patterns

    Journal of Chemical Neuroanatomy 16:77–116.

    https://doi.org/10.1016/s0891-0618(98)00065-9

    • PubMed
    • Google Scholar
18. 1. Kelly JG
    2. García-Marín V
    3. Rudy B
    4. Hawken MJ

    (2019) Densities and Laminar Distributions of Kv3.1b-, PV-, GABA-, and SMI-32-Immunoreactive Neurons in Macaque Area V1

    Cerebral Cortex 29:1921–1937.

    https://doi.org/10.1093/cercor/bhy072

    • PubMed
    • Google Scholar
19. 1. Kooijmans RN
    2. Sierhuis W
    3. Self MW
    4. Roelfsema PR

    (2020) A quantitative comparison of inhibitory interneuron size and distribution between mouse and macaque v1, using calcium-binding proteins

    Cerebral Cortex Communications 1:tgaa068.

    https://doi.org/10.1093/texcom/tgaa068

    • PubMed
    • Google Scholar
20. 1. Krienen FM
    2. Goldman M
    3. Zhang Q
    4. C H Del Rosario R
    5. Florio M
    6. Machold R
    7. Saunders A
    8. Levandowski K
    9. Zaniewski H
    10. Schuman B
    11. Wu C
    12. Lutservitz A
    13. Mullally CD
    14. Reed N
    15. Bien E
    16. Bortolin L
    17. Fernandez-Otero M
    18. Lin JD
    19. Wysoker A
    20. Nemesh J
    21. Kulp D
    22. Burns M
    23. Tkachev V
    24. Smith R
    25. Walsh CA
    26. Dimidschstein J
    27. Rudy B
    28. S Kean L
    29. Berretta S
    30. Fishell G
    31. Feng G
    32. McCarroll SA

    (2020) Innovations present in the primate interneuron repertoire

    Nature 586:262–269.

    https://doi.org/10.1038/s41586-020-2781-z

    • PubMed
    • Google Scholar
21. 1. Kubota Y
    2. Karube F
    3. Nomura M
    4. Kawaguchi Y

    (2016) The diversity of cortical inhibitory synapses

    Frontiers in Neural Circuits 10:27.

    https://doi.org/10.3389/fncir.2016.00027

    • PubMed
    • Google Scholar
22. 1. Lewis DA

    (2014) Inhibitory neurons in human cortical circuits: substrate for cognitive dysfunction in schizophrenia

    Current Opinion in Neurobiology 26:22–26.

    https://doi.org/10.1016/j.conb.2013.11.003

    • PubMed
    • Google Scholar
23. 1. Marín O

    (2012) Interneuron dysfunction in psychiatric disorders

    Nature Reviews. Neuroscience 13:107–120.

    https://doi.org/10.1038/nrn3155

    • PubMed
    • Google Scholar
24. 1. Medalla M
    2. Mo B
    3. Nasar R
    4. Zhou Y
    5. Park J
    6. Luebke JI

    (2023) Comparative features of calretinin, calbindin, and parvalbumin expressing interneurons in mouse and monkey primary visual and frontal cortices

    The Journal of Comparative Neurology 531:1934–1962.

    https://doi.org/10.1002/cne.25514

    • PubMed
    • Google Scholar
25. 1. Mehta P
    2. Kreeger L
    3. Wylie DC
    4. Pattadkal JJ
    5. Lusignan T
    6. Davis MJ
    7. Turi GF
    8. Li WK
    9. Whitmire MP
    10. Chen Y
    11. Kajs BL
    12. Seidemann E
    13. Priebe NJ
    14. Losonczy A
    15. Zemelman BV

    (2019) Functional access to neuron subclasses in rodent and primate forebrain

    Cell Reports 26:2818–2832.

    https://doi.org/10.1016/j.celrep.2019.02.011

    • PubMed
    • Google Scholar
26. 1. Mich JK
    2. Graybuck LT
    3. Hess EE
    4. Mahoney JT
    5. Kojima Y
    6. Ding Y
    7. Somasundaram S
    8. Miller JA
    9. Kalmbach BE
    10. Radaelli C
    11. Gore BB
    12. Weed N
    13. Omstead V
    14. Bishaw Y
    15. Shapovalova NV
    16. Martinez RA
    17. Fong O
    18. Yao S
    19. Mortrud M
    20. Chong P
    21. Loftus L
    22. Bertagnolli D
    23. Goldy J
    24. Casper T
    25. Dee N
    26. Opitz-Araya X
    27. Cetin A
    28. Smith KA
    29. Gwinn RP
    30. Cobbs C
    31. Ko AL
    32. Ojemann JG
    33. Keene CD
    34. Silbergeld DL
    35. Sunkin SM
    36. Gradinaru V
    37. Horwitz GD
    38. Zeng H
    39. Tasic B
    40. Lein ES
    41. Ting JT
    42. Levi BP

    (2021) Functional enhancer elements drive subclass-selective expression from mouse to primate neocortex

    Cell Reports 34:108754.

    https://doi.org/10.1016/j.celrep.2021.108754

    • PubMed
    • Google Scholar
27. 1. Mukherjee A
    2. Carvalho F
    3. Eliez S
    4. Caroni P

    (2019) Long-lasting rescue of network and cognitive dysfunction in a genetic schizophrenia model

    Cell 178:1387–1402.

    https://doi.org/10.1016/j.cell.2019.07.023

    • PubMed
    • Google Scholar
28. 1. Patiño M
    2. Lagos WN
    3. Patne NS
    4. Tasic B
    5. Zeng H
    6. Callaway EM

    (2022) Single-cell transcriptomic classification of rabies-infected cortical neurons

    PNAS 119:e2203677119.

    https://doi.org/10.1073/pnas.2203677119

    • PubMed
    • Google Scholar
29. 1. Prosniak M
    2. Hooper DC
    3. Dietzschold B
    4. Koprowski H

    (2001) Effect of rabies virus infection on gene expression in mouse brain

    PNAS 98:2758–2763.

    https://doi.org/10.1073/pnas.051630298

    • PubMed
    • Google Scholar
30. 1. Rudy B
    2. Fishell G
    3. Lee S
    4. Hjerling-Leffler J

    (2011) Three groups of interneurons account for nearly 100% of neocortical GABAergic neurons

    Developmental Neurobiology 71:45–61.

    https://doi.org/10.1002/dneu.20853

    • PubMed
    • Google Scholar
31. 1. Shapley R
    2. Hawken M
    3. Ringach DL

    (2003) Dynamics of orientation selectivity in the primary visual cortex and the importance of cortical inhibition

    Neuron 38:689–699.

    https://doi.org/10.1016/s0896-6273(03)00332-5

    • PubMed
    • Google Scholar
32. Book

    1. Shin H
    2. Adesnik H

    (2023) Functional roles of cortical inhibitory interneurons

    In: Usrey WM, editors. The Cerebral Cortex and Thalamus. Oxford University Press. pp. 72–80.

    https://doi.org/10.1093/med/9780197676158.003.0008

    • Google Scholar
33. 1. Tremblay R
    2. Lee S
    3. Rudy B

    (2016) GABAergic Interneurons in the neocortex: from cellular properties to circuits

    Neuron 91:260–292.

    https://doi.org/10.1016/j.neuron.2016.06.033

    • PubMed
    • Google Scholar
34. Conference

    1. Vafaei A
    2. Clark AM
    3. Federer F
    4. Angelucci A

    (2024)

    Computational function of parvalbumin-expressing inhibitory neurons in the primate primary visual cortex (V1)

    Society for Neuroscience 2024 Neuroscience Meeting Planner.

    • Google Scholar
35. 1. Van Brederode JFM
    2. Mulligan KA
    3. Hendrickson AE

    (1990) Calcium‐binding proteins as markers for subpopulations of GABAergic neurons in monkey striate cortex

    Journal of Comparative Neurology 298:1–22.

    https://doi.org/10.1002/cne.902980102

    • Google Scholar
36. 1. Verret L
    2. Mann EO
    3. Hang GB
    4. Barth AMI
    5. Cobos I
    6. Ho K
    7. Devidze N
    8. Masliah E
    9. Kreitzer AC
    10. Mody I
    11. Mucke L
    12. Palop JJ

    (2012) Inhibitory interneuron deficit links altered network activity and cognitive dysfunction in Alzheimer model

    Cell 149:708–721.

    https://doi.org/10.1016/j.cell.2012.02.046

    • PubMed
    • Google Scholar
37. 1. Vormstein-Schneider D
    2. Lin JD
    3. Pelkey KA
    4. Chittajallu R
    5. Guo B
    6. Arias-Garcia MA
    7. Allaway K
    8. Sakopoulos S
    9. Schneider G
    10. Stevenson O
    11. Vergara J
    12. Sharma J
    13. Zhang Q
    14. Franken TP
    15. Smith J
    16. Ibrahim LA
    17. Mastro KJ
    18. Sabri E
    19. Huang S
    20. Favuzzi E
    21. Burbridge T
    22. Xu Q
    23. Guo L
    24. Vogel I
    25. Sanchez V
    26. Saldi GA
    27. Gorissen BL
    28. Yuan X
    29. Zaghloul KA
    30. Devinsky O
    31. Sabatini BL
    32. Batista-Brito R
    33. Reynolds J
    34. Feng G
    35. Fu Z
    36. McBain CJ
    37. Fishell G
    38. Dimidschstein J

    (2020) Viral manipulation of functionally distinct interneurons in mice, non-human primates and humans

    Nature Neuroscience 23:1629–1636.

    https://doi.org/10.1038/s41593-020-0692-9

    • PubMed
    • Google Scholar
38. 1. Wood KC
    2. Blackwell JM
    3. Geffen MN

    (2017) Cortical inhibitory interneurons control sensory processing

    Current Opinion in Neurobiology 46:200–207.

    https://doi.org/10.1016/j.conb.2017.08.018

    • PubMed
    • Google Scholar
39. 1. Xu X
    2. Roby KD
    3. Callaway EM

    (2010) Immunochemical characterization of inhibitory mouse cortical neurons: three chemically distinct classes of inhibitory cells

    The Journal of Comparative Neurology 518:389–404.

    https://doi.org/10.1002/cne.22229

    • PubMed
    • Google Scholar
40. 1. Zhao P
    2. Zhao L
    3. Zhang T
    4. Qi Y
    5. Wang T
    6. Liu K
    7. Wang H
    8. Feng H
    9. Jin H
    10. Qin C
    11. Yang S
    12. Xia X

    (2011) Innate immune response gene expression profiles in central nervous system of mice infected with rabies virus

    Comparative Immunology, Microbiology and Infectious Diseases 34:503–512.

    https://doi.org/10.1016/j.cimid.2011.09.003

    • PubMed
    • Google Scholar

Author details

1. Frederick Federer

   Department of Ophthalmology and Visual Science, Moran Eye Institute, University of Utah, Salt Lake City, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1340-865X

2. Justin Balsor

   Department of Ophthalmology and Visual Science, Moran Eye Institute, University of Utah, Salt Lake City, United States

   Contribution

   Conceptualization, Formal analysis, Investigation, Methodology

   Competing interests

   No competing interests declared

3. Alexander Ingold

   Department of Ophthalmology and Visual Science, Moran Eye Institute, University of Utah, Salt Lake City, United States

   Present address

   Department of Electrical Engineering and Computer Science, University of Utah, Salt Lake City, United States

   Contribution

   Data curation, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0009-0007-5238-0016

4. David P Babcock

   Department of Ophthalmology and Visual Science, Moran Eye Institute, University of Utah, Salt Lake City, United States

   Present address

   Stritch School of Medicine, Loyola University, Chicago, United States

   Contribution

   Data curation

   Competing interests

   No competing interests declared

5. Jordane Dimidschstein

   Regel Therapeutics, Boston, United States

   Contribution

   Resources, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

6. Alessandra Angelucci

   Department of Ophthalmology and Visual Science, Moran Eye Institute, University of Utah, Salt Lake City, United States

   Contribution

   Formal analysis, Conceptualization, Supervision, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   alessandra.angelucci@hsc.utah.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1957-2231

• 626

  views

• 42

  downloads

• 0

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.